Metallic copper as a fiducial marker for both CT and PET

Patient positioning, immobilization devices, and fiducial markers in positron emission tomography-computed tomography scanner-based radiation therapy simulation

Positron emission tomography-computed tomography (PET-CT)-based radiation therapy (RT) treatment planning has become widely available in many RT practices. This broad acceptance has been facilitated by the RT specific features which are now commonly available on practically all modern PET-CT scanners. The quality and usability of PET-CT studies for RT treatment planning purposes strongly depends on patient positioning and immobilization during image acquisition. A well-developed clinical process and understanding of special needs for this type of imaging is crucial for successful use of PET-CT in radiation therapy planning and treatment. This article describes some of the important considerations in this process and provides several observations based on clinical experience with use of PET-CT imaging in radiation therapy planning and treatment. The use of fiducial markers in this process is also discussed.

PET/CT image registration: Preliminary tests for its application to clinical dosimetry in radiotherapy

The quality of dosimetry in radiotherapy treatment requires the accurate delimitation of the gross tumor volume. This can be achieved by complementing the anatomical detail provided by CT images through fusion with other imaging modalities that provide additional metabolic and physiological information. Therefore, use of multiple imaging modalities for radiotherapy treatment planning requires an accurate image registration method. This work describes tests carried out on a Discovery LS positron emission/computed tomography (PET/CT) system by General Electric Medical Systems (GEMS), for its later use to obtain images to delimit the target in radiotherapy treatment. Several phantoms have been used to verify image correlation, in combination with fiducial markers, which were used as a system of external landmarks. We analyzed the geometrical accuracy of two different fusion methods with the images obtained with these phantoms. We first studied the fusion method used by the PET/CT system by GEMS (hardware fusion) on the basis that there is satisfactory coincidence between the reconstruction centers in CT and PET systems; and secondly the fiducial fusion, a registration method, by means of least-squares fitting algorithm of a landmark points system. The study concluded with the verification of the centroid position of some phantom components in both imaging modalities. Centroids were estimated through a calculation similar to center-of-mass, weighted by the value of the CT number and the uptake intensity in PET. The mean deviations found for the hardware fusion method were: deltax/ +/-sigma = 3.3 mm +/- 1.0 mm and /deltax/ +/-sigma = 3.6 mm +/- 1.0 mm. These values were substantially improved upon applying fiducial fusion based on external landmark points: /deltax/ +/-sigma = 0.7 mm +/- 0.8 mm and /deltax/ +/-sigma = 0.3 mm 1.7 mm. We also noted that differences found for each of the fusion methods were similar for both the axial and helical CT image acquisition protocols.