Room: Exhibit Hall | Forum 6
Purpose: To present a novel phantom designed for the characterization and optimization of dual-energy (DE) imaging parameters for use with an on-board imager (OBI).
Methods: A phantom was designed and built to evaluate bone suppression capabilities using a DE imaging system with an OBI. The phantom was constructed of 16 cm lung-equivalent material sandwiched between 2 cm tissue-equivalent slabs. Five simulated tumors (0.5 â€“ 2.5 cm) located at two different depths were encased in the lung-equivalent materials. Two slabs with bone-equivalent material inserts were constructed to simulate ribs, which overlap the simulated tumors. The phantom was used to evaluate different combinations of energy pairs (high: 100-140 kVp; low 50-90 kVp) using the OBI of a commercial linear accelerator. Bone suppression was performed using a weighted logarithmic subtraction of the energy pairs and characterized using the contrast-to-noise ratio (CNR). Noise was evaluated using the noise power spectrum (NPS). Based on the NPS, different noise reduction techniques were evaluated on the resulting DE images.
Results: For all simulated tumors, our analysis demonstrated that the 140-60 kVp pair produced the highest CNR of 2.94 Â± 0.19 following bone subtraction. Moreover, the top two noise reducing algorithms were auto encoders with Convolutional-Neural-Networks (CNN) architecture and the Simple Smoothing Method (SSM) applied to the resulting DE images. Application of noise reduction techniques increased CNR values by more than twice to 6.86 Â± 0.65.
Conclusion: We present a novel phantom designed to characterize and optimize parameters associated with DE imaging. This phantom allows for quick image acquisition, optimization of imaging parameters and weighting factors. It is expected that the DE images with increased soft tissue visibility will enhance automated lung tumor tracking allowing for real-time adaptive radiotherapy.
Funding Support, Disclosures, and Conflict of Interest: Supported by NIH R01CA207483.