Evaluation of ultra-short focusing cone-beam collimators for SPECT brain imaging

We evaluated short focal length cone-beam collimators for a dual-head brain SPECT system. Previously we showed that using a 20-cm-focal length cone-beam collimator paired with a 40-cm-focal length fan-beam collimator increases sensitivity dramatically near the center of the brain. In this study, we evaluated the signal-to-noise ratio (SNR) for various lesion locations when using short focusing collimators of different focal length. A spherical lesion was placed at each of several possible locations in an ellipsoidal phantom. The focal length was varied from 18 to 23 cm, so that the focal point was always inside the brain for a 15-cm radius of rotation. We calculated the SNR for estimation of activity concentration based on the Cramer-Rao lower bound on variance. For centrally located lesions, the SNR gain compared to a low-energy-high-resolution collimator ranged from 2.5 (f = 23 cm) to 3.6 (f = 18 cm). For peripheral lesions, the gain ranged from 1.7 to 1.8. We also reconstructed images of the Zubal brain phantom for simulated /sup 123/I-Altropane studies. Images were reconstructed from noisy projection data using the OS-EM algorithm. Based on image comparisons and SNR calculations, we concluded that the combination of a cone-beam collimator with focal point inside the brain and a fan-beam collimator (f = 40 cm), outperforms any other conventional collimators in sensitivity. Furthermore, mismatch between the center of the patient's head and the center of rotation was found not to reduce the sensitivity significantly.