Impacto do contraste no cálculo do standardized uptake value (SUV) em função do tamanho da lesão e distância de áreas de maior concentração de contraste

Abstract

It has been shown that the use of contrast media in PET/CT (positron emission tomography/ computed tomography) studies is beneficial when compared to an unenhanced CT scan. This is due to the fact that contrast media generates better contrast regarding adjacent structures, allows better characterization of abnormalities, and also allows visualizing alterations that do not present abnormal metabolic behavior. However, contrast agents generate more absorption of CT photons in relation to those of PET potentially giving rise to errors in attenuation correction of PET images. Errors converting the attenuation map from the CT to the PET scan may lead to inaccurate quantification of PET images. The occurrence of SUV (Standardized uptake value) modifications in lesions when PET images are corrected with either contrast-enhanced or unenhanced CT scans has already been studied. Our goal, however, was to study the occurrence of such alterations in relation with lesion size, and distance from the lesion to areas of high contrast concentration, such as high caliber vases and focal areas of the gastrointestinal tract (GIT). Methods: We obtained the SUVmax of 149 lesions, from 26 patients who underwent a PET/CT scan in which images were corrected using both an unenhanced computed tomography (PETCTUE) and a contrast-enhanced computed tomography (PETCTCE). In addition, we measured lesion sizes (smallest and largest diameter, and area) and their distance to large vessels and the GIT. Size measures were not taken when the lesion could not be visualized in the CT image. For lesions in which the distance to high contrast concentration areas was too large or too small, and could not be accurately estimated, the median distance of similar lesions was used to replace the measures. Following the repeated measures principle, we obtained Lin's concordance correlation coefficient and its confidence interval. A Bland-Altman analysis was performed using the absolute difference values and ratio, between PETCTCE and PETCTUE with concordance limits. Results: A total of 149 lesions could be visualized in PET. Eleven of them could not be identified in the CT, and so their measurements were not taken. For 16 lesions, the distance to high contrast concentration areas could not be obtained. The concordance level between the methods, by Lin's concordance correlation coefficient, was 0.99 (CCC: 0.99), and the confidence interval was 95% of 0.98 – 0.99 (IC95%: 0.98 – 0.99). The mean absolute difference between the methods was approximately zero and the relative difference was +3.37% (concordance limits of 95% between -2.72 and +2.72, and between -15.73 and +22.48%, respectively). Eighty-two lesions had their larger diameter below 17mm and the other 56 had a larger diameter above or equal to 17mm, with mean percent variation of SUVmax from PETCTCE to PECCTUE of 3.85% and 2.83%, respectively (p=0.54). Sixty-six lesions had a smaller diameter below 17mm and the remaining 72 had a smaller diameter equal to or above 17mm. The mean percent SUVmax variation from PETCTCE to PECCTUE was 2.71%±1,74 and 3.80% ±1,60, respectively (p=0.41). Eighty-two lesions had an area larger than 2.27 cm2 and the other 57 had an area equal to or larger than 2.27 cm2. The mean percent variation of the SUVmax obtained for PETCTCE and PETCTUE was 2.81%±1.64 e 3,83%±1.70, respectively (p=0.48). Sixty-seven lesions were more than 1.82 cm distant from large vessels/ GIT, and 82 were at a less than 1cm distance. The mean SUVmax variation between the methods (PETCTCE versus PETCTUE) was 0.57±1.65% e 4.98±1.61% respectively (p < 0.005). Conclusions: For the totality of studied lesions (149), as well as for the analysis regarding the factors size (non significant p) and distance (significant p), the medium SUVmax variation we identified it not clinically relevant. Thus, both contrast-enhanced and unenhanced PET/CT can be used for attenuation correction.