PET/CT imaging fusion combines the precise anatomical description provided by CT (a) with the ability to detect the uptake activity of tumor cells, as we can see in (b). Thus, using PET/CT image fusion (c), we have at one’s disposal more information than with each image modality separately, which is very useful when delimiting GTV in radiotherapy treatment.
Side view of the flangeless PET esser phantom (a) used to analyze fusion under attenuation and scatter conditions similar to those found on scanning a patient. It contains a set of hollow spheres with diameters ranging from following a distribution as scheme shown in (b) so that the PET axial image should be obtained as displayed in (c).
Frontal (a) and side (b) views of the tube phantom, designed to be able to obtain a large number of slices that would provide us with reliable statistics to do a geometric evaluation of the image fusion. It consisted of a distribution of length tubes placed in a poliespan test tube rack and it was surrounded by a set of fiducials. Fiducials were made up of a PMMA panel with catheters inside it. Both tubes and fiducial catheters were filled with a mixture of CT and PET contrast agents so that the phantom PET reconstructed frontal view (MIP) should be obtained as displayed in (c).
CT and PET images obtained for the phantoms and axial image resulting from the hardware fusion. In the tube phantom (a), as well as in the cylindrical phantom (b), we can see in the detailed enlarged portions, results from the hardware fusion are poor as the superimposed structures are clearly displaced.
Graphic representation of the distribution of the relative shift of tube centroids, for both types of fusion and with both CT imaging modalities (axial and helical). (a) shows displacements on the axis, whereas (b) shows displacement on the axis. Fiducial fusion gives the best results in all cases, with mean values of around . No significant difference can be observed between axial or helical image acquisition modes.
Draft of the fiducial frame designed in our department for its regular use in acquiring PET/CT images of virtual simulation for radiotherapy treatment. In the outline has been pointed out fillable fiducial catheters (a) and the laser centering mark (b).
Correspondence in scales in axes and main diagonals in CT and PET axial. Differences between CT and PET images are less than .
As shown by the values, relative rotation between both reconstruction centers is almost negligible, although it is possible to see a significant shift between them.
Summary of the mean results and standard deviations obtained in the structures evaluated for both fusion modalities. The left side shows the values obtained with the tube phantom, which allows us to evaluate a large number of points. The right side shows the results for the phantom under attenuation and scatter conditions similar to those a patient would undergo. As one can observe in both cases, displacement obtained by hardware fusion can be corrected through the semi-automatic fusion based on external fiducial points, providing mean values are below .
Article metrics loading...
Full text loading...