The feasibility of using cone beam computed tomography (CBCT) images for Monte Carlo (MC) brachytherapy dose calculations has been investigated. To evaluate the effects of tissue heterogeneities and finite patient dimensions for 192Ir high dose rate treatment, CT-based MC calculations for breast and head and neck cases were first performed using the PTRAN_CT photon transport code. PTRAN_CT is an accelerated MC code specifically designed for patient-specific dose calculations. Muscles and adipose tissues, which are nearly indistinguishable in CBCT images, are found to cause minimal dose perturbations at 192Ir energies compared to water. The proximity of the tumor to the skin, however, will have an observable impact on the dose up to a few percent. Therefore, for CBCT calculations, a reasonable assignment of material and density values to the patient voxel geometry, with a good delineation of the skin and bony structures, will suffice for MC dose calculations. A CBCT-based calculation for an actual treatment plan with the tumor close to the cheek was performed. The results were compared to TG43 calculations to quantify the dose differences in the target and critical structures. Since the dose delivered to the tumor is mostly primary dose, deviations are found mostly in the organs at risk where scatter contribution becomes more significant. This study shows that for HDR brachytherapy applications, CBCT-based MC calculations is a feasible option despite inferior image quality and larger uncertainties in the Hounsfield Units compared to CT images. Research supported by Nucletron BV.
Keywords: Brachytherapy; Cancer; Computed tomography; Cone beam computed tomography; Dosimetry; Medical image quality; Medical imaging; Monte Carlo methods; Skin; Tissues.
© 2008 American Association of Physicists in Medicine.