2416 - CBCT-Based Dosimetric Comparison of Two Generic Imaging Algorithms

J. L. Smetanick¹, E. L. Clouser¹, B. Laughlin², C. Kutyreff¹, Y. Rong¹, and R. C. Tegtmeier¹; ¹Department of Radiation Oncology, Mayo Clinic AZ, Phoenix, AZ, ²Department of Radiation Oncology, Mayo Clinic, Phoenix, AZ

Purpose/Objective(s): This investigation compared the dosimetric performance of cone-beam CT (CBCT)-based photon (PH) and proton (PR) intact prostate plans generated on both iterative CBCT (iCBCT) and corrected CBCT (corrCBCT) datasets. The ultimate aim of this comparison was to evaluate utilization of commercially available enhanced CBCT imaging techniques to monitor accuracy in daily dose delivery for PH and PR radiotherapy. Materials/

Methods: Standard PH and PR intact prostate stereotactic body radiotherapy plans (3625 cGy in 5 fractions) were retrospectively generated on deformed planning CT (pCT) images for 25 intermediate-risk patients with no nodal or metastatic involvement. The PH volumetric modulated arc therapy plans consisted of two full treatment arcs with 10 MV flattening filter-free beams and planning target volume margins of 3mm (2mm posteriorly for rectal sparing). PR scanning beam single-field uniform dose plans were generated with two lateral fields and optimization target volume margins of 3mm (8mm laterally for distal end uncertainty). For each patient, plans were propagated to segmented final fraction CBCT images reconstructed with each technique (the only difference between the datasets being Hounsfield Unit (HU) values). To quantify HU variation for each approach relative to pCT, the HU mean absolute error (MAE) was computed. To evaluate dosimetric performance, recalculated PH and PR doses were compared with 3D gamma analysis for various tolerances and for dose-volume histogram (DVH) metrics according to the RTOG0938 protocol. Statistical analysis of DVH metrics was performed via a two-sided t-test (for an alpha of 0.05).
Results: The HU MAE relative to pCT was reduced by over a factor of 4 for corrCBCT compared to iCBCT (5.4±4.1 HU vs. 23.8±19.8 HU, respectively). When dosimetrically comparing techniques, mean gamma pass rates at 3%/3mm (with a 10% dose threshold) were 99.9±0.2 (PH) and 99.9±0.1 (PR), while at 1%/1mm passing rates were slightly reduced to 99.8±0.1 and 99.8±0.2, respectively. All target volume DVH metrics (D0.03cc, D95%, Dmin) were within 1.0% on average for both PH and PR plans (with only the variation in Dmin being statistically significant). Discrepancies between bladder and rectum metrics (D1cc, D90%, D50%) were within 0.5% for PH plans. This variation was increased for PR (up to 2.6% for rectum D90%) and these differences were generally statistically significant (though not clinically significant when considering actual metric values). No statistical variation was observed in femoral head metrics for either plan type.
Conclusion: While HU accuracy relative to pCT was noticeably improved for corrCBCT compared to iCBCT, analysis of both gamma pass rates and DVH metrics indicated a similar dosimetric performance between the two reconstruction/post-processing techniques for both plan types. The results of this study highlight the utility of each commercially available imaging solution for accurate CBCT-based dose evaluation in PH and PR radiotherapy.