A delta-scatter Monte Carlo model of a rotating parallel-hole collimator for /sup 131/I brain tumor imaging

We investigate the collimator response and sensitivity of a lead rotating parallel-hole collimator (RPHC), 6.4 cm thick, with 1.45 mm hexagonal holes and 2.0 mm septa, designed to image ¹³¹I emissions (364.5 keV, 82%). The collimator is rotated to each of three positions during imaging to give data for three independent, offset hole arrays, filling the septal shadows in each array with data from the other two arrays. Collimator response can be incorporated into the reconstruction algorithm to improve image quality. Simple geometric calculations of collimator response do not account for Compton scatter. For ¹³¹I, the effects of downscatter into the photopeak from high-energy emissions (637 keV, 6.5%, and 723 keV, 1.7%) are not obvious. A Monte Carlo model incorporating delta scattering is used to investigate downscatter and primary transmission. For initial studies, a perfectly absorbing detector is used. The RPHC is modelled as plate a 6.4 cm thick, with photon path lengths chosen randomly using the attenuation coefficient of lead. For interaction points determined to be in holes, photons are allowed to continue unimpeded (delta scatter). This increases the number of path lengths generated, but greatly decreases the number of boundary intersection calculations. Point sources are modelled from 2.5 cm to 25 cm from the collimator both on and off of hole axes. At short distances, photons are seen through one hole or a small number of holes. As distance increases, changes in the response and sensitivity are clearly seen. Downscatter presents little problem, constituting less than 10% of the detected counts except when the point source is close to the collimator, and between holes, i.e., in the septal shadows. Data for these areas will be filled in as the collimator is moved to other rotational positions