Abstract
485
Objectives In the application of gating in imaging there is an inherent trade-off between information gained from motion control and lost from decreased signal, so that it can be difficult for the human eye to decipher the cost/benefit of using the gating. To address this we developed a method which utilizes the information in the 4D data to optimize its presentation. This algorithm, termed “gating+”, is fast, robust, and fully automated.
Methods Our method is centered on frequency-based analysis of information at the voxel level. Specifically, a single set of data is gated twice:A) with accurate gating triggers, and B) with randomly placed triggers. The frequencies present in each voxel within the volume are then weighted between the voxel’s DC value and its gated fluctuation using the ratios of the amplitudes of the max random frequency and the 1 period/cycle accurately gated frequency. We applied our gating+ algorithm to basic lesion and human phantom scan simulations, as well as small animal μPET data.
Results Gating+ images were compared to gated and non-gated images, voxel-specific standard deviation maps were generated for simulations. Processing reconstructed μPET volume required <20 s. Gating+ images consistently exhibited resolution improvements similar to that of gated scans, while maintaining lower noise levels in uniform or low count areas. These measures are exemplified for a lesion simulation in our table,and are typical of what we observed in μPET data. Total activity was conserved in the gating+ images, relative noise varied.
Conclusions The current practice of viewing a single gated image disregards useful information acquired in the other n-1 gates of the scan, information about patient motion and real signal that our algorithm utilizes. The algorithm works robustly with post-reconstructed “effective” signal. This approach is elegant (<15 lines high level code) can be integrated into black box software with other motion compensation (gating and motion mapping) strategies.
Research Support USIEF Fulbrigh