| Summary: | Background. In the study of 3D images of the brain according to emission computed tomography (SPECT) there is a problem of their spatial orientation for the correct construction of sections and subsequent analysis of the symmetry of effective hemisphere perfusion. The brain on the original images can take virtually any orientation, resulting in asymmetry of distribution of radiopharmaceutical agents (RPA) on sections of parallel or perpendicular coordinate axes. In turn, this can lead to incorrect interpretation of clinical data.
Objective. The purpose of the paper is development of a methodology and appropriate software module for the automated standardization of spatial orientation of the SPECT of the brain images, in particular, with 99mTc-hexamethylpropylene (99mTc-HMPAO) perfusion RPA.
Methods. 30 SPECT images of patients with different levels of brain perfusion were analyzed. Brain scintigraphic studies were performed with 99mTc- HMPAO on a gamma camera "E. Cam"(Siemens) with LEHR collimator. SPECT was performed in 15–20 min after intravenous administration of RPA. The tomographic study included the collection of 128 projections for the 128×128 matrix, the RPA activity for intravenous administration was 740 MBq. The developed method of automated standardization of spatial orientation SPECT image of the brain was implemented in the software "ScintyBrain" in Matlab 2018.
Results. The developed technique, based on finding the plane of symmetry of the hemispheres of the brain, allows determining the spatial orientation of the brain with an average error of 0.8–4°. The error in determining the angle of rotation of the sagittal sections increases with decreased perfusion of the frontal segments, and the frontal sections – depends on the presence of large-focal changes in the temporal and parietal divisions of the hemisphere.
Conclusions. A method of automated spatial standardization of SPECT images of the brain was developed. The average error of the spatial angles of rotation of 3D images is in the range of 0.8–4°. The average computer analysis of the brain's spatial orientation takes up to 40 sec. The implementation of the presented method minimizes the subjective influence of specialists on the results of processing and analysis of scintigraphic images. In addition, the automated SPECT standardization procedure can significantly reduce the time of subsequent blending of SPECT brain images with other modalities.
|
|---|