Uncertainties in calculating the geometric parameters of radiopharmaceutical accumulation lesion on diagnostic images during pre-radiation preparation for radiosurgerye

The phantom (imitating the brain) with different diameter models-spheres installed inside was scanned on positron emission tomography (PET) combined with computed X-ray tomography (CT) and magnetic resonance tomography (MRI). When MR scans of this phantom were performed to visualize the instilled spheres, their volumes were filled with the contrast agent copper sulfate pentahydrate (CuSO4.5H2O), which contributes to its sensitivity to chemical shifts and susceptibility artifacts (additional causes of distortion arising from encountering density differences in diagnostic MRI and radiation therapy planning). Qualitative assessment of the obtained images for each sequence for artifacts was performed. Quantitative analysis of the obtained PET/CT scanning results was performed, namely, the accuracy of recovery of volumetric activity of model-spheres (pathological foci of radiopharmaceutical accumulation). Taking into account high practical significance of treatment of pathological formations of small volume, spheres with inner diameter of 4 and 6 mm were used additionally as models (pathological foci). The influence of the number of subsets on the diagnostic informativity of PET/CT images, as well as on the value of recovery coefficients was established. Contouring of models-spheres on CT and MR images by five radiologists was performed in order to evaluate the influence of personal variability on the final contour of structures. Also, models of pathologic foci were delineated on the obtained PET/CT images of different modalities, and differences in nominal values and delineation volumes were calculated. Deviation values of contrast recovery coefficients of modelsphere models with diameters of 4–22 mm on PET images reconstructed with different parameters were obtained. The influence of reconstruction parameters on the formation of the resulting volume of model-sphere delineation of different diameter on PET images reconstructed with different number of subsets (18 and 24), as well as with different values of the Gaussian filter was evaluated.