Program Information
Comparison of Integrated and Post-Reconstruction Dual-Energy CT Proton Stopping Power Ratio Estimation Approaches
S Zhang1*, D Han2 , D Politte3 , J Williamson3 , J O'Sullivan1 , (1) Washington University in St. Louis, St. Louis, MO, (2) Virginia Commonwealth University Medical Center, Richmond, VA, (3) Washington University School of Medicine, St. Louis, MO
Presentations
SU-K-605-5 (Sunday, July 30, 2017) 4:00 PM - 6:00 PM Room: 605
Purpose: This work compares the proton stopping power ratio (SPR) estimation accuracy of an integrated dual-energy CT (DECT) approach, which combines dual-energy image reconstruction and proton SPR mapping under the same framework, with well-accepted post-reconstruction DECT approaches that use two independently reconstructed single-energy images.
Methods: Simulated DECT transmission data (90/140 kVp) of two 330 mm cylindrical water phantoms containing 34 standard human tissue inserts with ICRU-recommended compositions were used to evaluate the achievable accuracies of three different DECT SPR estimation approaches: (a) an integrated method based on a basis vector model (BVM) in which two BVM component images used for SPR estimation are reconstructed simultaneously using our dual-energy statistical image reconstruction (SIR) algorithm; (b) a post-reconstruction BVM approach in which two CT images are reconstructed separately using a mono-energetic SIR algorithm followed by calibration using an additional phantom; and (c) a post-reconstruction Torikoshi parametric fitting model (tPFM) approach which uses the reconstructed images from (b) and estimates the SPR via a spectrum-averaged tPFM with additional calibration. In addition to full sized phantoms, we also use 20% smaller phantoms to investigate the effect of patient size change on estimation accuracy.
Results: The root-mean-square (RMS) SPR errors for 34 ICRU tissues were 0.27%, 0.65% and 0.54% for the methods (a), (b), and (c), respectively. The corresponding standard deviations were 1.14%, 2.69% and 3.06%, respectively. Decreasing the phantom diameter by 20% did not affect the accuracy of the integrated method, but increased the RMS errors for the two post-reconstruction approaches by 0.30% and 0.23%.
Conclusion: The integrated DECT approach based on statistical dual-energy image reconstruction outperforms the SPR mapping accuracy of post-reconstruction approaches, and is much less vulnerable to CT imaging uncertainties. A systematic experimental evaluation under clinical condition is warranted.
Funding Support, Disclosures, and Conflict of Interest: Funding Support: NCI R01 CA 149305, NCI R01 CA 212638; No Financial Disclosures; No Conflicts of Interest.
Contact Email: