Generating the synthetic CT (sCT) and synthetic MR (sMR: sT1w/sT2w) images of the brain using atlas based method

Abstract

<strong>Introduction</strong>: Radiation therapy planning (RTP) is one of the clinical applications in which both CT scan and MRI are used. MR and CT images are applied to determine the target volume and calculation of dose distribution, respectively. In addition, using two imaging modalities increases the department workload and cost. In this study, an algorithm was presented to create synthetic CT (sCT) and synthetic MR (sMR: sT1w/sT2w) images consequently CT or MR images are not taken from patient. <strong> </strong> <strong>Materials and Methods:</strong> The synthetic CT (sCT) and synthetic MR (sMR: sT1w/sT2w) images was generated by atlas based method. The dataset consisted of MR and CT paired images from 17 brain RT patients. To generate sCT/sMR image, first each MR/CT atlas was registered to the MR/CT target image, the resulting transformation was applied to the corresponding CT/MR atlas which created the set of deformed images. Then, deformed images were fused to generate a single sCT/sMR. The sCT/sMR was compared with the real CT/MR images using the MAE. <strong>Results:</strong> The evaluations performed by MAE showed a good agreement between the sCT/sMR and real CT/MR images. The results represented 72±11.3 HU, 51±7.3 image intensity and 64±7.3 image intensity for sCT, sT1w and sT2w, respectively. <strong> </strong> <strong>Conclusion:</strong> This method showed that sT1w images have lesser MAE with respect sT2w. As a result, CT in transferring the brain tissues geometry to sT1w is more successful than sT2w. Also, the sCT images based T1w are in better agreement with real CT than sCT images based T2w.