Effect of filtering on the detection and localization of small Ga-67 lesions in thoracic single photon emission computed tomography images

Tumor detection can be significantly affected by filtering so determining a n optimal filter is an important aspect of establishing a clinical reconstr uction protocol, The purpose of this study was to identify the cue-off freq uency of a Butterworth filter used in a filtered backprojection (FBP) recon struction that maximized the detection and localization accuracy of 1 cm sp herical lesions in Ga-67 citrate, thoracic SPECT images. Image quality was evaluated by means of a localization receiver operating characteristic (LRO C) study using computer simulated images. Projection data were generated us ing the mathematical cardiac-torso digital phantom with a clinically realis tic background source distribution. The images were reconstructed using FBP with multiplicative Chang attenuation correction and fifth-order Butterwor th filtering. The cut-off frequencies considered were 0.25, 0.32, 0.47, and 0.79 cm(-1) for the case of three-dimensional (3D) post-filtering and 0.25 , 0.32, and 0.47 cm(-1) for two-dimensional (2D) post-filtering. The images were read by three research scientists and one board certified nuclear med icine clinician.. The area under the LROC curve and the localization accura cy for all test conditions were compared using Scheffe's multiple compariso ns test. It was found that 3D post-filtering using filters with cut-off fre quencies of 0.32 and 0.47 cm(-1) resulted in the highest lesion detectabili ty and localization accuracy. These two test conditions did not differ sign ificantly from each other but were significantly better (p<0.05) than all o f the 2D, and the 3D 0.79 cm(-1) cut-off frequency cases. (C) 1999 American Association of Physicists in Medicine. [S0094-2405(99)01007-X].