Estudo da estratégia de harmonização da quantificação do SUV em imagens de PET/CT

Abstract

In clinical practice, PET/CT images are often analyzed qualitatively by visual comparison of tumor lesions and normal tissues uptake; and semi-quantitatively by means of a parameter called SUV (Standardized Uptake Value). To ensure that longitudinal studies acquired on different scanners are interchangeable, and information of quantification is comparable, it is necessary to establish a strategy to harmonize the quantification of SUV. The aim of this study is to evaluate the strategy to harmonize the quantification of PET/CT images, performed with different scanner models and manufacturers. For this purpose, a survey of the technical characteristics of equipment and acquisition protocols of clinical images of different services of PET/CT in the state of Rio Grande do Sul was conducted. For each scanner, the accuracy of SUV quantification, and the Recovery Coefficient (RC) curves were determined, using the reconstruction parameters clinically relevant and available. From these data, harmonized performance specifications among the evaluated scanners were identified, as well as the algorithm that produces, for each one, the most accurate quantification. Finally, the most appropriate reconstruction parameters to harmonize the SUV quantification in each scanner, either regionally or internationally were identified. It was found that the RC values of the analyzed scanners proved to be overestimated by up to 38%, particularly for objects larger than 17mm. These results demonstrate the need for further optimization, through the reconstruction parameters modification, and even the change of the reconstruction algorithm used in each scanner. It was observed that there is a decoupling between the best image for PET/CT qualitative analysis and the best image for quantification studies. Thus, the choice of reconstruction method should be tied to the purpose of the PET/CT study in question, since the same reconstruction algorithm is not adequate, in one scanner, for qualitative and quantitative assessments in different scopes. We concluded that the harmonization strategy of the SUV quantification presented in this paper was effective in reducing the variability of small structures quantification. However, for the comparison of SUV quantification between different scanners and institutions, it is essential that, in addition to the harmonization of quantification, the standardization of the methodology of patient preparation is maintained, in order to minimize the SUV variability due to biological factors.