mm). This represented a significant decrease in the error in image fusion at the location of the markers of 48% (P < 0.001). Figure 1(b) shows a fused image after pixel size correction. Discussion The suspected relative difference in real image size between PET and CT was confirmed using both phantom studies and clinical scans. In the clinical scans, a series of separate measurements over multiple patients demonstrated the deviation with statistical significance. The manual identifica- tion of the centre of the markers was excluded as a possible source of error in the analysis procedure, because the inter- operator variability was negligible. The results of the phantom measurements and the clinical scans were concordant (the size differences in the phantom measurements were well within the standard deviations of the clinical series). The deviation was shown to be linear, and thus the results from the phantom study and the clinical data were theoretically exchange- able. The image size difference derived from the clinical series was considered to be the most accurate, because statistical analysis of a series of measurements is less prone to bias in the manual evaluation of markers than a single measurement in the phantom experiment. There- fore, the image size difference derived from the clinical series was applied as the correction factor. This approach was validated by repeated measurements after correction, which no longer demonstrated significant image size differences. The observed difference in image size between PET and CT was larger in the transverse plane than in the axial direction (2.0% versus 0.8%). The exact cause of this discrepancy remains unclear. There is no apparent reason why image size calibration of PET images would be more difficult in one direction than in another. The relative difference in real image size between PET and CT was caused by an absolute error in calibration of the PET image size. With a pixel size and effective resolution (full-width at half-maximum) in the range of 5 mm in PET imaging, accurate image size calibration may be difficult. For example, to detect a deviation of 1%, the difference to be found in a marker distance over a length of 20 cm is only 2 mm, which is well below the resolution. Therefore, the procedure to determine the image size of a PET scanner will always be less accurate relative to a CTscanner. Furthermore, in IMRT planning, CT must be the gold standard by default because planning of the radiation fiel...
mm). The accrual of interest and the accretion or amortization of original issue discount on Indebtedness and the payment of interest in the form of additional Indebtedness originally incurred in accordance with this Section 6.01 will not constitute an incurrence of Indebtedness.
mm). The Contractor shall be bound by this DPA for the duration of the Service Agreement or so long as the Contractor maintains any Student Data.
