Dosimetric verification of synthetic CT using Cone Beam CT in an MRI only workflow

Sammanfattning: Background and purpose: The MRI only workflow aims to base the treatment planning solely on magnetic resonance imaging (MRI), hence excluding the traditional computed tomography (CT) scan. To do that synthetic computed tomography (sCT) data is generated, replacing the conventional CT data. The aim of this study was to investigate the possibility of using cone beam computed tomography (CBCT) images to verify the sCT data. This was done via investigation of the properties of the kV CBCT systems in the clinic and comparing dose distributions carried out on both the sCT and the CBCT data. Materials and methods: Several phantom measurements were made on the kV CBCT systems on the six Varian TrueBeamTM linac used in the clinic. The properties investigated were the variation in Hounsfield Units (HUs) over time for one kV CBCT system as well as variation between the six kV CBCT systems. The HUs of one kV CBCT system was also compared to HUs from a Siemens CT system. Using 28 CBCT data sets from seven patients, a HU to relative electron density (RED) table was created. Treatment plans (RapidArc) were calculated on sCT, CT and CBCT images using the standard HU-RED table based on CT HU. For the CBCT data, additional calculation was done using the HU-RED table based on CBCT HU. The difference between the dose distributions was evaluated comparing clinical dose volume histograms (DVH) and mean absorbed doses. Results: The phantom measurements showed that the kV CBCT system was stable in HU over time. All six kV CBCT systems generated comparable HU values. The variation of HUs between CT and CBCT images was minor. The CBCT images, however, exhibited larger variation across the field of view (FOV) compared to CT images. Dose calculation based on CBCT data showed a mean dose difference to PTV of 0.0% (HU-RED CT) and -0.8% (HU-RED CBCT) compared to dose calculations based on sCT data. Conclusion: The HUs for one kV CBCT system were found to stable over time. The variation between the six kV CBCT systems were found to be minor. Results obtained for one system can therefore be transferred onto all the systems. The minor difference in HUs between CT and CBCT images indicates that a correction of HUs is not necessary to obtain sufficiently accurate absorbed dose calculations on CBCT images. The compared dose distributions based on CBCT and sCT data showed good agreement in terms of dose accuracy, regardless of which HU-RED table used. This indicates that CBCT data can be used to verify sCT data.