Purpose: X-ray Computed Tomography (CT) has been an exceptionally powerful and versatile tool in diagnostic imaging and radiotherapy. Despite numerous technological evolutions, its underlying mechanism for image formation remains unchanged. In this study, we introduce a new image formation approach, termed pair production tomography (P2T) imaging, to expand the capabilities of X-ray tomography.
Methods: Different from CT which utilizes X-ray transmission signals, P2T collects coincident annihilation photons originated from the pair production interaction with high energy MV X-rays for tomographic reconstruction. This study proposes three P2T acquisition and reconstruction methods, including filtered back projection (FBP), time-of-flight (TOF) assuming a high time-resolution detector, and scanning pencil beam (SPB) where the imaging FOV is excited sequentially. The feasibility of P2T imaging was tested using Monte Carlo simulation on phantom and patient data.
Results: Three distinctive capabilities of P2T were demonstrated: in vivo 3D radiotherapy dosimetry verification and monitoring, the ability to form tomography with truncated-view and as few as a single X-ray beam, and high linearity with the material atomic number for element mapping and soft tissue differentiation. Among the three P2T acquisition methods, FBP is the most straightforward to achieve engineeringly, but its utility is limited to high radiation dose procedures such as radiotherapy treatment monitoring. Both TOF and SPB result in high signal-to-noise ratio (SNR) P2T images with a typical imaging dose. The image quality of TOF relies on the time resolution of detectors. In comparison, SPB inherently suppresses localization errors by triangulating the signal source based on the known excitation path, thus achieving high SNR comparable to that using ultra-fast detectors.
Conclusion: The novel P2T imaging expands the capabilities of X-ray tomography and is achievable with existing engineering technologies.
Funding Support, Disclosures, and Conflict of Interest: This research is supported by DOE Grants Nos. DE-SC0017057 and DE-SC0017687, NIH Grants Nos. R01CA188300, R43CA183390, and R44CA183390.