increasing interests in using magnetic resonance imaging only in radiation therapy require method... more increasing interests in using magnetic resonance imaging only in radiation therapy require methods for predicting the computed tomography numbers from MRi data. Here we propose a simple voxel method to generate the pseudo-ct (pct) image using dual-contrast pelvic MRi data. the method is first trained with the CT data and dual-contrast MRI data (two sets of MRI with different sequences) of multiple patients, where the anatomical structures in the images after deformable image registration are segmented into several regions, and after MRi intensity normalizations a regression analysis is used to determine a two-variable polynomial function for each region to relate a voxel's two MRi intensity values to its CT number. We first evaluate the accuracy via the Hounsfield unit (HU) difference between the pseudo-CT and reference-CT (rCT) images and obtain the average mean absolute error as 40.3 ± 2.9 HU from leave-one-out-cross-validation (LOOCV) across all six patients, which is better than most previous results and comparable to another study using the more complicated atlas-based method. We also perform a dosimetric evaluation of the treatment plans based on pCT and rCT images and find the average passing rate within 2% dose difference to be 95.4% in point-to-point dose comparisons. therefore, our method shows encouraging results in predicting the ct numbers. this polynomial method needs less computer storage than the interpolation method and can be readily extended to the case of more than two MRi sequences. The workflow of conventional radiation therapy (RT) uses computed tomography (CT) images to create the treatment plan and to position the patient at treatment. Magnetic resonance imaging (MRI) is also often used to provide precise delineation of RT target volumes due to its superior soft tissue contrast 1. When these two modalities are both used, the workflow then necessitates an additional step of registering the images of the two modalities: MRI and CT 2. Recently, the concept of MRI-Linac using magnetic resonance imaging only in radiation therapy, i.e., MRI-only RT, has become more popular, because MRI-only RT has the benefits of improving the workflow and removing systematic errors in registering MRI and CT images 3,4. However, in MRI-only RT there is a problem in creating treatment plans with the MRI images because of the lack of CT images or the electron density information. To address this problem, various methods of generating CT images, called pseudo-CT (pCT), have been investigated 5-7. Existing methods in deriving a pseudo-CT from MR images may be classified into different categories. They include the classification into segmentation-based, intensity-based, atlas-based, and hybrid methods 7 , or the classification into segmentation-, atlas-, patch-, and learning-based methods 8 , or the classification into bulk density assignment, atlas-based, and voxel-based methods 9. These methods have produced mean absolute error (MAE) values ranging from 85 HU 10 to 137 HU 11 for the brain and from 36.5 HU 12 to 74.3 HU 13 for the prostate (pelvis region), for example. Among these methods, atlas-based methods 12,14-17 align an MRI atlas, which has been derived from an MRI database pre-registered to the corresponding reference-CT (rCT) images, to the target patient's MRI images through registration. The atlas thus contains predetermined correlations between the MRI voxels and the variables of interest such as the CT number or organ type. The same registration (with the translational, rotational, and deformable information) that maps the MRI atlas to the target patient's MRI images is then applied to the atlas CT images to create the target patient' pCT images. This approach is popular because of its potential in producing
increasing interests in using magnetic resonance imaging only in radiation therapy require method... more increasing interests in using magnetic resonance imaging only in radiation therapy require methods for predicting the computed tomography numbers from MRi data. Here we propose a simple voxel method to generate the pseudo-ct (pct) image using dual-contrast pelvic MRi data. the method is first trained with the CT data and dual-contrast MRI data (two sets of MRI with different sequences) of multiple patients, where the anatomical structures in the images after deformable image registration are segmented into several regions, and after MRi intensity normalizations a regression analysis is used to determine a two-variable polynomial function for each region to relate a voxel's two MRi intensity values to its CT number. We first evaluate the accuracy via the Hounsfield unit (HU) difference between the pseudo-CT and reference-CT (rCT) images and obtain the average mean absolute error as 40.3 ± 2.9 HU from leave-one-out-cross-validation (LOOCV) across all six patients, which is better than most previous results and comparable to another study using the more complicated atlas-based method. We also perform a dosimetric evaluation of the treatment plans based on pCT and rCT images and find the average passing rate within 2% dose difference to be 95.4% in point-to-point dose comparisons. therefore, our method shows encouraging results in predicting the ct numbers. this polynomial method needs less computer storage than the interpolation method and can be readily extended to the case of more than two MRi sequences. The workflow of conventional radiation therapy (RT) uses computed tomography (CT) images to create the treatment plan and to position the patient at treatment. Magnetic resonance imaging (MRI) is also often used to provide precise delineation of RT target volumes due to its superior soft tissue contrast 1. When these two modalities are both used, the workflow then necessitates an additional step of registering the images of the two modalities: MRI and CT 2. Recently, the concept of MRI-Linac using magnetic resonance imaging only in radiation therapy, i.e., MRI-only RT, has become more popular, because MRI-only RT has the benefits of improving the workflow and removing systematic errors in registering MRI and CT images 3,4. However, in MRI-only RT there is a problem in creating treatment plans with the MRI images because of the lack of CT images or the electron density information. To address this problem, various methods of generating CT images, called pseudo-CT (pCT), have been investigated 5-7. Existing methods in deriving a pseudo-CT from MR images may be classified into different categories. They include the classification into segmentation-based, intensity-based, atlas-based, and hybrid methods 7 , or the classification into segmentation-, atlas-, patch-, and learning-based methods 8 , or the classification into bulk density assignment, atlas-based, and voxel-based methods 9. These methods have produced mean absolute error (MAE) values ranging from 85 HU 10 to 137 HU 11 for the brain and from 36.5 HU 12 to 74.3 HU 13 for the prostate (pelvis region), for example. Among these methods, atlas-based methods 12,14-17 align an MRI atlas, which has been derived from an MRI database pre-registered to the corresponding reference-CT (rCT) images, to the target patient's MRI images through registration. The atlas thus contains predetermined correlations between the MRI voxels and the variables of interest such as the CT number or organ type. The same registration (with the translational, rotational, and deformable information) that maps the MRI atlas to the target patient's MRI images is then applied to the atlas CT images to create the target patient' pCT images. This approach is popular because of its potential in producing
Uploads
Papers by Samuel Leu