364 - Improving SPR Calculation for Pediatric Patients Using a Novel Imaging Algorithm for Tissue Parameterization with Photon-Counting CT
Presenter(s)
J. Zhu1, K. Qing1, B. Liu1, X. Wang2, D. Zhang3, C. Shi4, S. Penfold5, T. M. Williams1, and A. Liu1; 1Department of Radiation Oncology, City of Hope National Medical Center, Duarte, CA, 2Rutgers Cancer Institute of New Jersey, New Brunswick, NJ, 3University of California, Los Angeles, Los Angeles, CA, 4Department of Radiation Oncology, Orange County Lennar Foundation Cancer Hospital, Irvine, CA, 5Australian Bragg Centre for Proton Therapy and Research, Adelaide, SA, Australia
Purpose/Objective(s): Tissue parameterization is a novel technique for determining the chemical composition of tissues from CT images. This approach has the potential to identify tissue types based on the weights of their chemical components. Applications of tissue parameterization include patient-specific stopping power ratio (SPR) calculations and enhancements in imaging contrast and tissue identification using photon-counting CT (PCCT).
Materials/Methods: PCCT can generate monoenergetic CT images at different kV energy levels. The monoenergetic CT number represents the fractionated and accumulated attenuation coefficients of each chemical element composition. Simulated monoenergetic CT images were generated at energy levels ranging from 40 keV to 140 keV in 5 keV increments. A phantom containing 11 types of pediatric tissues was used, with the mass attenuation coefficients obtained from the NIST XCOM database.
Since the weight of each element in a pixel remains consistent across energy levels, a system of equations was formulated for each pixel based on varying beam energies. The Diagonally Relaxed Orthogonal Projections (DROP) algorithm, a block-iterative method related to component averaging, was used to solve these equations and determine the elemental weights. The Bragg additivity rule was then applied, allowing for SPR calculation based on electron density.Results: The calculated elemental weights were compared to the reference data. The deviations for all tissues remain within ±1% for hydrogen (H), sulfur (S), chlorine (Cl), potassium (K), calcium (Ca), and iron (Fe). Deviations for nitrogen (N) and sodium (Na) are under 2%. Oxygen (O) exhibits a maximum deviation of 6.3% from reference values for one tissue; however, given that its absolute weight always exceeds 60%, this deviation is considered reasonable. The SPR values were also calculated and compared with the references in Table 1, with absolute differences remaining within 0.01, except for cortical bone (0.014).
Conclusion: This study demonstrates the feasibility of determining the elemental composition of each pixel using PCCT. Unlike traditional methods that rely on CT number-to-SPR calibration curves, this approach directly calculates tissue composition for the first time. By leveraging elemental weights, this method shows promise for tissue identification and classification.
Abstract 364 - Table 1: The comparison of different SPRs in pediatric human tissues| reference SPR | measured SPR | deviation | |
| Adipose | 0.995 | 1.000 | 0.005 |
| Brain(Infant) | 1.034 | 1.039 | 0.005 |
| Heart(Child-4-18y) | 1.042 | 1.047 | 0.005 |
| Kidney(Child 2y) | 1.043 | 1.048 | 0.005 |
| Liver(Child 1y) | 1.050 | 1.056 | 0.006 |
| Lung(Fetus 17-40 weeks) | 1.043 | 1.048 | 0.006 |
| Muscle(Infant 3 months) | 1.050 | 1.056 | 0.006 |
| Cortical Bone(Child 15y) | 1.648 | 1.662 | 0.014 |
| Femur(Child 15y) | 1.258 | 1.264 | 0.006 |