Data Elements

Leadership:

• (co-leader) Adam L. Kesner, Memorial Sloan Kettering Cancer Center, New York, NY, USA
• (co-leader) Kris Thielemans, University College London, London, UK

The ETSI’s “Data Elements” sub-group is responsible for developing a scheme for the emission tomography raw data models to enable a hierarchical description of the various scanner architectures via a full multi-layered geometric format consisting of basic detector “modules”, together with a list of rigid transformations to replicate the “modules” across the multiple layers of the architecture. The scheme would describe data structures, for example list mode streams, at full fidelity (i.e. “uncompressed” or span=1, no view/TOF mashing etc.), by creating a list of detector-element-index-pairs, and optionally TOF-bin and /or energy-window-pairs.

Therefore, normalization and dead-time factors for full-fidelity data would be required by the format. Full-size “full-fidelity” data  however can become quite large, which can be an important consideration for long axial field of view (LAFOV) scanners. Due to these considerations, storing full-fidelity normalization/dead-time factors might currently not be desirable, especially as we need to be able record the time evolution of the dead-time. We therefore propose to rely on simple component-based models.

The proposed model also accounts for concurrent physiologic signal definitions.

The developing emission tomography raw data standard and associated software development kit (SDK) to sustain the use of the data formats can be found at ETSI’s GitHub public repository.