The effect of point source configuration on MHR/COR calibration.

Objectives Center of rotation (COR) correction is a calibration and quality control procedure performed and applied to align gamma camera image centers to the gantry’s physical center of rotation. Values obtained from the calibration are used to correct clinically acquired SPECT projections prior to reconstruction. Without these properly applied corrections, points of activity in the object being imaged would reconstruct into rings, distorting the visualized source distribution. The Siemens method, termed Multiple Head Registration (MHR), uses an array of five point sources arranged in a provided source holder. As the name implies, the procedure also calculates and applies a correction in the Y-axis to align each detector to one another in the axial direction of the patient. Siemens provides vials with a conical base for the creation of point sources when performing this calibration. Per Siemens specifications, the volume of the source would be less than 0.1-0.15 mL. The aim of this project is to compare the results of MHR corrections acquired using extremely small volume point sources (as close as possible to ideal) to corrections acquired using sources in a less than ideal configuration (i.e. activity spotted paper towel placed in source vials).

Methods Idealized point sources were produced by placing approximately two drops of Tc-99m solution into the conical portion of the point source vials with a 25g needle. Each vial contained between 1079 and 1150 uCi. Five additional point sources were created by placing a small section of paper towel into each vial and adding approximately five to six drops of Tc-99m solution with a 25g needle. Each vial contained between 954 and 1066 uCi. Each set of point sources was imaged on four Siemens SPECT systems: two E.cams, a Symbia and an Intevo. MHRs (COR and Y-shift) were calculated and recorded. The E.cam systems calculate a specific X-shift for each projection angle and characterize the calibration by indicating the Min X and Max X-shift (pixels) for each detector. A Y-shift value is also calculated. The Symbia and Intevo systems calculate and characterize the calibration by providing a single value for COR (X-shift) for each detector, as well as the Y-shift.

Results For the E.cam systems, the magnitude of the X-shift is determined by the absolute value of Max X - Min X (pixels). Across the four detectors (two for each camera) the average increase in absolute X-shift, generated using the less than ideal sources, was 0.011 pixels equalling only a 0.054 mm increase. The increase in Y-shift was 0.004 pixels or 0.019 mm. For the Symbia and Intevo systems, the calibrated COR shift values were greater for the idealized sources compared to the less than ideal sources by an average of 0.057 mm, excluding a single value of 0.052 mm for the Intevo’s detector two. The idealized sources’ Y-shift also increased, but only by 0.009 mm.

Conclusions It was hypothesized that the less than ideal point sources would create inaccuracies in the MHR determination, therefor creating larger MHR correction values. The findings of this project indicate negligible differences in MHR correction values when acquired with less than ideal versus idealized point sources.