Quantifying the Shift Patterns Off bone in Fiducial-Based IG-SBRT for Pancreatic Adenocarcinoma

For patients undergoing radiation for pancreatic adenocarcinoma (PDAC), daily variation in target positioning, primarily a consequence of respiratory motion and bowel gas patterns, presents a challenge to highly conformal treatment. While on-board imaging, such as cone-beam computed tomography (CBCT), and endoscopically-placed fiducials allow for daily shifts to optimize target positioning, the magnitude of shift may directly impact target coverage and dose to organs-at-risk (OARs). Herein, we aim to characterize the daily shifts required for patients treated with stereotactic body radiation therapy (SBRT) for PDAC at a high-volume institution. We reviewed 30 consecutive patients with PDAC who underwent SBRT using alpha-cradle immobilization, active breathing control (ABC), and daily CBCT. Patients were initially aligned to bone then subsequently shifted to align to fiducials. The last CBCT taken prior to treatment initiation was reviewed, and Velocity software was used to quantify the shift from bone to fiducials in the Superior-Inferior (SI), Left-Right (LR), and Anterior-Posterior (AP) axes. The magnitude of the shift was also calculated as a three-dimensional (3D) vector. The mean, maximum, and standard deviation (SD) of the shift in each axis, including the 3D vector, were calculated for individual patients over five fractions. Population mean and SD of each metric was also computed. Of note, the SD of the shift across individual patients was used as a surrogate for inter-fraction variation. Finally, to understand if a large day one shift predicted greater variability in shifts throughout treatment, a paired T-test was performed to assess the association between the mean shift and the inter-fraction variation in the shift to fiducial. The mean shift to fiducial in the L-R, A-P, and S-I axes was 3.25 ± 1.88 mm, 2.74 ± 1.80 mm, and 3.83 ± 2.81 mm, respectively, and the mean 3D vector shift was 6.59 ± 3.28 mm. On average, the maximum shift to fiducial in the L-R, A-P, and S-I axes was 5.73 ± 2.77 mm, 4.87 ± 2.50 mm, and 7.08 ± 4.2 mm, respectively, and the maximum 3D vector shift was 9.58 ± 4.30 mm. The inter-fraction variation in shifts was 1.89 ± 1.00 mm, 1.65 ± 0.88 mm, and 2.37 ± 1.45 mm in the L-R, A-P, and S-I axes, while the inter-fraction variation for the 3D vector shift was 4.06 ± 1.51 mm. On paired T-test analysis, increased shifts in all the axes and increased 3D vector shifts were significantly associated with more inter-fraction variation in the shift to fiducial. For patients undergoing pancreatic SBRT, daily variations in target position require clinically significant shifts that must be commonly applied in all three axes, which may impact daily dose to the target and OARs. A large shift to fiducial on day one predicts continued large shifts throughout treatment course. Efforts to account for such variation are critical for successful delivery of conformal radiation.