Purpose: In conventional log-subtraction soft-tissue dual-energy (DE) x-ray systems, bone is suppressed by selecting a constant weighting factor across the image. However, tumor contrast also depends on the weighting factor. The purpose of this work is to develop a DE imaging system to enhance tumor contrast by assigning pixel-based weighting factors in the regions of the tumor while different weighting factors to suppress the bone elsewhere.
Methods: A step phantom constructed containing slabs of soft tissue (solid water) in one direction and bone material in the other direction with the thickness ranges of [0,30] and [0,6] cm respectively forming 7x7 regions. Two tumor plates were constructed with tumor cylinders (height 2 or 4 cm) for each 7x7 region. Using a clinical system, phantom x-ray images were acquired at 60 and 140 kVp. For each bone overlap region, optimal weighting factors were calculated by either minimizing bone contrast-to-noise-ratio (wBmin) or maximizing tumor contrast-to-noise-ratio (wTmax).
Results: Due to beam hardening effect, both wBmin and wTmax varied depending on the amount of soft-tissue or bone thickness. The wBmin varied in the range [0.61, 1.12] and its value was higher for thicker bone or soft-tissue regions. The wTmax had a clear peak for each region, was lower than wBmin for each corresponding region, and varied in the range [0.37, 0.68]. Minimal variations in wTmax were observed between the 2 and 4 cm thick tumor plates.
Conclusion: This work demonstrates that using a priori CT data, pre-calculated weighting factors can be assigned to each pixel to create a DE image such that tumor contrast is maximized in the regions of tumor while bone contrast is minimized elsewhere. For image-guided radiation therapy (IGRT) applications, such a priori CT is always available from simulation CT and the tumor location can also be calculated from digitally reconstructed radiographs.
Funding Support, Disclosures, and Conflict of Interest: Atlantic Canada Opportunities Agency, Atlantic Innovation Fund