Delivery Verification of Tomotherapy Treatments Using Exit Detector Sinograms: A Phantom Study
V Rodriguez1,2*, S Richardson2, T Ghosh1, R Kashani2, S Goddu2, (1) University of Missouri-Columbia, Columbia, MO, (2) Washington University School of Medicine, St. Louis, MO,SU-E-T-182 Sunday 3:00:00 PM - 6:00:00 PM Room: Exhibit Hall
The objective of this study was to use a liquid water female body phantom to investigate the feasibility of using helical Tomotherapy's exit detector data sinograms (EDDS) for quality assurance of Tomotherapy treatments.
Delivery Verification software (DV-Tool), described by Kapatoes et al., uses EDDS and daily mega-voltage-CT images for delivery sinogram reconstruction and subsequent dose computations. These dose distributions should reflect variations in patient's anatomy and machine delivery. Phantom was used to verify the robustness of the DV-Tool for machine consistency over 2months, and sensitivity to weight loss and setup errors. Weight loss simulations were done by removing layers of bolus from the phantom's pelvis. Setup errors were simulated by performing known table shifts before the treatment. Two treatment plans simulating a breast and gynecological treatment plans were used for these evaluations. Reconstructed doses and dose volume histograms (DVHs) were analyzed in comparison to the plan DVHs by scaling the reconstructed PTV mean dose (PTV-MD) to match the plan. PTV-MD differences before and after machine output adjustments were evaluated.
Machine consistency check of gynecological plan revealed that the deviations were greater (±5%) during the week prior to target replacement. After output adjustments, the deviations for breast and gynecological plans agreed within ±1.0% and ±2.5%, respectively. Weight loss on the gynecological plan showed a correlation (R2=0.998) between lost thickness and increasing PTV-MD by approximately 5%/cm. Setup errors on the breast plan showed broadening shoulder of the PTV DVH for shifts greater than 5mm. Volume at 20Gy for the right lung increased by 6% for 10mm shifts in both x and z directions.
The DV-Tool is sensitive to simulated weight loss, setup errors, and machine output. Exit detector data and the DV-Tool have great potential for monitoring patient's delivered treatments. However, clinical workflow needs to be established.