OBJECTIVE: We present a validation study of an algorithm for retrospective registration of PET and MR brain images. MATERIALS AND METHODS: This algorithm involves two steps. In the first step, the two volumes are reformatted by aligning their interhemispheric fissure planes (midsagittal plane). In the second step, the corresponding planes parallel to the midsagittal plane are further aligned in the reformatted volumes to produce a 3D rigid body registration of the two original volumes. It is an efficient algorithm because both steps are performed in 2D spaces, and in each step only a small number of landmarks are required. A user-friendly system has been implemented to facilitate easy and fast processing of registration and reformatting of image volumes. The accuracy of this algorithm is validated using clinical scans of neurosurgical patients with a stereotaxic frame attached to their skull. The frame-based stereotaxic system provides an effective method for transforming image coordinates from different image volumes into a common coordinate system. This common coordinate system is used for assessing the spatial correspondence of each pixel in the registered image volumes. Validation using the stereotaxic image volumes enables objective estimation of retrospective registration accuracy. RESULTS: Analysis of 11 MR/PET image pairs indicates that our registration method not only is efficient but also provides adequate accuracy for most clinical evaluation of PET studies. CONCLUSION: We have implemented and validated an efficient algorithm for retrospective registration of PET and MR brain images.
OBJECTIVE: We present a validation study of an algorithm for retrospective registration of PET and MR brain images. MATERIALS AND METHODS: This algorithm involves two steps. In the first step, the two volumes are reformatted by aligning their interhemispheric fissure planes (midsagittal plane). In the second step, the corresponding planes parallel to the midsagittal plane are further aligned in the reformatted volumes to produce a 3D rigid body registration of the two original volumes. It is an efficient algorithm because both steps are performed in 2D spaces, and in each step only a small number of landmarks are required. A user-friendly system has been implemented to facilitate easy and fast processing of registration and reformatting of image volumes. The accuracy of this algorithm is validated using clinical scans of neurosurgical patients with a stereotaxic frame attached to their skull. The frame-based stereotaxic system provides an effective method for transforming image coordinates from different image volumes into a common coordinate system. This common coordinate system is used for assessing the spatial correspondence of each pixel in the registered image volumes. Validation using the stereotaxic image volumes enables objective estimation of retrospective registration accuracy. RESULTS: Analysis of 11 MR/PET image pairs indicates that our registration method not only is efficient but also provides adequate accuracy for most clinical evaluation of PET studies. CONCLUSION: We have implemented and validated an efficient algorithm for retrospective registration of PET and MR brain images.
Authors: Zhaolin Chen; Sharna D Jamadar; Shenpeng Li; Francesco Sforazzini; Jakub Baran; Nicholas Ferris; Nadim Jon Shah; Gary F Egan Journal: Hum Brain Mapp Date: 2018-08-04 Impact factor: 5.038
Authors: Thomas Funck; Kevin Larcher; Paule-Joanne Toussaint; Alan C Evans; Alexander Thiel Journal: Front Neuroinform Date: 2018-09-26 Impact factor: 4.081