Room: Track 1
Purpose: Due to the distinct radiosensitivity of proliferating hematopoietic stem cells in active bone marrow, identifying and sparing active marrow may significantly improve radiotherapy treatment. ¹8F-fluoro-l-deoxythymidine positron emission tomography (FLT-PET) differentiates active marrow from fatty marrow. However, FLT-PET is not FDA approved and cannot be used repeatedly for image-guided radiotherapy. We propose a multi-material decomposition (MMD) framework to perform bone marrow decomposition using only dual-energy CT (DECT) images.
Methods: The proposed MMD was formulated as an optimization problem including a quadratic data fidelity term, along with an isotropic total variation term to regulate image smoothness, and a non-convex penalty function to promote sparsity in the decomposition domain. The mass and volume conservation rule were formulated as the probability simplex constraint. The DE ratio of calcium was included in the data fidelity term to separate calcium in the marrow. The optimization problem was solved using an accelerated primal-dual splitting approach with line search for non-convex problems. The volume fraction of active marrow was evaluated using the proposed DECT decomposition method on a swine dataset and was compared with FLT-PET/MR and water-fat MRI scans of the same animal.
Results: The proposed algorithm decomposed bone marrow into active marrow, fatty marrow, and calcium. The volume fractions of active marrow computed from DECT accord with that of PET, with vertebrae and sacrum ala having most active marrow and femur having most fatty marrow. The volume fraction of active marrow evaluated from [MRI, DECT, PET] are [0.20, 0.04, 0.05], [0.16,0.14,0.17], [0.51, 0.78, 0.54], and [0.51, 0.85, 0.36] for femur, ilium, sacrum ala, and vertebrae, respectively. The correlation is 0.90 between PET and MR, and 0.92 between PET and DECT.
Conclusion: The proposed bone marrow decomposition method on DECT images identifies active bone marrow that is consistent with the FLT-PET and water-fat MRI.
Funding Support, Disclosures, and Conflict of Interest: This work is supported by NIH Grants Nos. R01CA188300, R43CA183390, and R44CA183390, and DOE Grants Nos. DE-SC0017057 and DE-SC0017687.