The aim of this study was to evaluate the effect of the CT-derived attenuation correction on ^99mTc-MIBI normal myocardial perfusion scintigrams. Rest perfusion scintigrams of patients in whom coronary artery disease was suspected, without a history or any signs in ECG of a myocardial infarction, were analysed. Patients were included in the material if their rest perfusion scintigrams were normal. This criterion was fulfilled by 61 patients (29 men and 32 women) aged between 40 and 74 (mean value 57) years, with body mass between 50 and 120 (mean value 70) kg. Tomographic reconstruction of a radionuclide study was performed with an iterative OSEM method (10 subsets, 2 iterations) sequentially without and with attenuation and scatter corrections on a dedicated Xeleris workstation, applying an ACQC tool to enable manual realignment of SPECT and CT images. SPECT studies were evaluated visually and semiquantitatively. Visual analysis of tomograms was performed with the aim of finding sites of significantly lower counts in comparison with the maximal level (in the lateral wall). Semiquantitative analysis was based on counts in 20 segments of a polar map. Attenuation correction caused a complete (in 32 of 40 - 80% of patients) or partial (in 8 of 40 - 20% of patients) filling out of all areas of lower counts in the inferior wall. However, although in the anterior wall attenuation correction caused a complete (in 11 of 35 - 31% of cases) or partial (10 of 35 - 29% of cases) filling of areas of lower counts, in 14 cases (40%) those areas remained unchanged or increased, and in 8 cases (13% of all patients) new areas of decreased counts appeared. The same was true for the apical region, in which areas of decreased counts were detected in 14 of 61 (23%) cases without attenuation correction, but after application of the correction number of apical defects, this figure grew to 22 (36%) patients. Altogether, attenuation correction reduced the total number of lower count areas from 104 to 66. Semi-quantitative analysis revealed that attenuation correction reduced nonuniformity in counts in the whole myocardium - the mean difference between segment with maximum counts and values in all segments was reduced from 17.5 ± ± 12% to 11.0 ± 10.3% (p < 0.0001) in male patients, and in female patients, from 11.5 ± 9% to 10.5 ± 8.6%, thus equalling non-uniformities in myocardial scintigrams of both sexes. Misalignment of CT and SPECT studies was observed in 17 (28%) patients but only in 2 (3% of all patients) patients did CT realignment evidently change the attenuation corrected scintigrams.
Although attenuation correction can cause artefacts, its use is justified by the reduction of the total number of areas of lower counts and the improvement of uniformity of images of normally perfused myocardium.