The HURRA filter: An easy method to eliminate collimator artifacts in high-energy gamma camera images

• Published in: Revista Española de medicina nuclear e imagen molecular (English ed.) 2017-01, Vol.36 (1), p.27-36
• Main Authors: Perez-Garcia, H, Barquero, R
• Format: Article
• Language: English
• Subjects: 131I therapy; 2D gamma-camera image; Activity quantification; Artefacto en la imagen de gammacámara; Colimador de alta energía; Corrector filter; Cuantificación de la actividad; Filtro corrector; Fourier transform; Gamma camera image artifact; High-energy collimator; Imagen de gammacámara 2D; Radiology; Terapia con 131I; Transformada de Fourier
• ISSN: 2253-8089; 2253-8089
• DOI: 10.1016/j.remnie.2016.06.007

Abstract Purpose The correct determination and delineation of tumor/organ size is crucial in 2-D imaging in131 I therapy. These images are usually obtained using a system composed of a Gamma camera and high-energy collimator, although the system can produce artifacts in the image. This article analyses these artifacts and describes a correction filter that can eliminate those collimator artifacts. Methods Using free software, ImageJ, a central profile in the image is obtained and analyzed. Two components can be seen in the fluctuation of the profile: one associated with the stochastic nature of the radiation, plus electronic noise and the other periodically across the position in space due to the collimator. These frequencies are analytically obtained and compared with the frequencies in the Fourier transform of the profile. A specially developed filter removes the artifacts in the 2D Fourier transform of the DICOM image. This filter is tested using a 15-cm-diameter Petri dish with131 I radioactive water (big object size) image, a131 I clinical pill (small object size) image, and an image of the remainder of the lesion of two patients treated with 3.7 GBq (100 mCi), and 4.44 GBq (120 mCi) of131 I, respectively, after thyroidectomy. Results The artifact is due to the hexagonal periodic structure of the collimator. The use of the filter on large-sized images reduces the fluctuation by 5.8–3.5%. In small-sized images, the FWHM can be determined in the filtered image, while this is impossible in the unfiltered image. The definition of tumor boundary and the visualization of the activity distribution inside patient lesions improve drastically when the filter is applied to the corresponding images obtained with HE gamma camera. Conclusion The HURRA filter removes the artifact of high-energy collimator artifacts in planar images obtained with a Gamma camera without reducing the image resolution. It can be applied in any study of patient quantification because the number of counts remains invariant. The filter makes possible the definition and delimitation of small uptakes, such as those presented in treatments with131 I.