An evaluation of the effect of partial-septa on detection of small lesions in brain phantom study using MDAPET camera

We investigated the effect of partial-septa on detection of small lesions embedded into the Hoffman brain phantom using a high resolution PET camera. Three-dimensional (3D) positron emission tomography (PET) acquisition in comparison to two-dimensional (2D) PET acquisition improves the sensitivity of the system at the cost of higher random and scatter coincidence contributions. A partial-septa system, in which full ring septa are inserted between only some of the detector rings, allowing "partial" 3D acquisition may provide a better alternative than the 2D acquisition and full 3D acquisition for reducing the scatter and random coincidence with less sensitivity loss. We used the MDAPET camera to detect small lesions in brain images obtained from the scan of the 3D Hoffman brain phantom with three embedded small lesion phantoms with diameters of 3, 5, and 8.6 mm. To include the effect of background from activity outside the FOV, a large cylindrical uniform phantom with diameter of 21.5 cm filled with lower activity was placed near the Hoffman brain phantom. Data were acquired in three separate scans for 3 phantoms: (1) the large uniform cylindrical phantom; (2) the Hoffman brain phantom; (3) the 3 small lesions embedded in brain phantom. For scan of each phantom, the other phantoms were filled only with water. For lesions, data were collected for several standard uptake values (SUVs) as activity decayed. The three sets of data were selectively combined to generate the sinograms for the desired SUVs. In partial-septa configurations, the white matter regions of the brain images had lower counts and the images visually looked better than no-septa images. Images taken in the partial-septa configuration had slightly higher contrast