Measurements of recovery coefficients for the HR+ and their application in a correction method

Introduction
The activity concentration of small lesions with diameters comparable to the spatial resolution (FWHM) is underestimated in PET measurements. This resolution is determined by the camera characteristics and also by the reconstruction and filter conditions used. For simple formed lesions the underestimation can be corrected, if the size of the lesion and the spatial resolving in the recon-structed picture are known and appropriate phantom measurements are present.
The aim of the study was to determine the resolution as well as the recovery coefficients( HSRC) for small spheres by phantom measurements and to prove, whether the underestimation is possible to correct by means of the quotient of the two measured lesion values for reconstructions with clearly different picture resolution.

Methods
All measurements took place at the PET camera ECAT EXACT HR+ with the software V7.1 (Siemens, CTI) in the 2D-Mode. For the measurement of the HSRC the EEC head phantom with the hollow spheres insert was used. Additionaly 8 thin-walled glass spheres with inside diameter values between 5.0 mm and 8.6 mm were examined. The reconstruction of the phantom measurements took place with FBP and OSEM.
The transversal resolution was measured with three line sources in water and the average values from radial and tangential values were calculated.
From ROIs with the diameter equal to the FWHM the recovery coefficients (phantom measurements) and the contrast values (patient measurements) were determined.

Results and Discussion
The transver-sal resolution was measured for axial distances up to 11.5 cm. For distances larger than 7 cm the radial component is increased.
If the HSRC from two reconstructions with significantly different image resolution are used to calculate a quotient HSRC1/HSRC2 then spheres with diameters smaller approximatly 15 mm can be distinct.
This method can be used to correct the recovery of a lesion in a patient study by determination of the image contrast. Fore the application of the procedure to lesions in the periphery of the body the dependence of the resolution from the axial distance must be considered.
Further investigation should be done to show the limits of this method.