Room: Exhibit Hall | Forum 9
Purpose: The accuracy of range prediction is a limiting factor in plan robustness for proton therapy. This study investigates the accuracy of range predictions based on stopping power ratio (SPR) images constructed from dual-energy CT (DECT).
Methods: DECT scans were acquired on a Siemens SOMOTON Definition Edge CT scanner, with 26 tissue surrogate plugs inserted individually into a high-density polyethylene phantom. SPR images were created for each tissue plug using Ï?e and Zeff images constructed from the DECT scans. For each plug, the calculated SPR value was compared with the measurement performed using a multilayer ionization chamber (MLIC). To further validate the range prediction of SPR images, a water-phantom setup was designed and DECT and single-energy CT (SECT) scans were performed for each plug. The SPR images created by DECT scans were imported into RayStation treatment planning systems (TPS) for direct dose measurement. The SECT image created using a standard stoichiometric calibration curve were imported into TPS. Plans were created on the SPR image and recalculated on the SECT image. Plans for Lung-Exhale, HDPE, and Bone-200 plugs were irradiated and compared with measured dose at multiple depths (proximal, mid, and distal locations in the target). Measurements were compared with the TPS dose calculated based on SPR and SECT images.
Results: The DECT derived SPR values agreed with the measured values to within 2.0% for all tissue surrogates except muscle, liver (<2.5% each) and lung (~16%). For the measurements in water-phantom setup, the calculated dose on SPR image showed better agreement to the ion chamber measured dose (in the target: Bone-200 and HDPE within 3% and Lung-Exhale within 5%) than SECT image.
Conclusion: Our study suggests that in comparison to using SECT images, the SPR images constructed from DECT can improve the accuracy of range prediction especially for non-natural materials.
Not Applicable / None Entered.