Summary

The aim of this study was to evaluate the anatomical accuracyand reproducibility of retrospective interactive rigidimage registration (RIR) between routinely archived X-raycomputer tomography (CT) and positron emission tomographyperformed with 18F-deoxyglucose (FDG-PET) inoncological patients. Methods: Two observers registeredPET and CT data obtained in 37 patients using a commerciallyavailable image fusion tool. RIR was performed separatelyfor the thorax and the abdomen using physiologicalFDG uptake in several organs as a reference. One observerperformed the procedure twice (O1a and O1b), anotherperson once (O2). For 94 malignant lesions, clearly visiblein CT and PET, the signed and absolute distances betweentheir representation on PET and CT were measured in X-, Y-,and Z-direction with reference to a coordinate system centeredin the CT representation of each lesion (X-, Y-, Z-distances).Results: The mean differences of the signed andabsolute distances between O1a, O1b, and O2 did not exceed3 mm in any dimension. The absolute X-, Y-, andZ-distances ranged between 0.57 ± 0.58 cm for O1a(X-direction) and 1.12 ± 1.28 cm for O2 (Z-direction).When averaging the absolute distances measured by O1a,O1b, and O2, the percentage of lesions misregistered byless than 1.5 cm was 91 % for the X-, 88 % for the Y-, and77 % for the Z-direction. The larger error of fusion determinedfor the remaining lesions was caused by non-rigidbody transformations due to differences in breathing, armposition, or bowel movements between the two examinations.Mixed effects analysis of the signed and absolute X-,Y-, and Z-distances disclosed a significantly greater misalignmentin the thorax than in the abdomen as well asaxially than transaxially. Conclusion: The anatomical inaccuracyof RIR can be expected to be <1.5 cm for the majorityof neoplastic foci. Errors of alignment are bigger in thethorax and in Z-direction, due to non-rigid body transformationscaused, e.g., by breathing.