Influence of Field of View and Bowtie Filtration on Cone Beam Computed Tomography Image Quality and...

The image quality (IQ) of cone-beam computed tomography (CBCT) is often reduced due to X-ray scatter, causing issues such as shading, skin-line artifacts, decreased contrast-to-noise ratio, and inaccurate computed tomography (CT) numbers. This study establishes six metrics for assessing IQ, focusing on both traditional metrics, such as contrast-to-noise ratio, and clinically relevant measures of CT signal accuracy. Using a commercial CBCT system for image-guided radiation therapy (IGRT), the study examines how these metrics vary with axial field-of-view (FOV_z) and bowtie filter use to understand the effects of X-ray scatter on IQ.

Catphan-600 phantom was scanned at five longitudinal FOV_z settings (2-27 cm, Superior-Inferior) with and without a bowtie filter, and all software-based scatter corrections were disabled. Six metrics were evaluated: shading (m_shading), periphery accuracy (m_periphery), noise (m_noise), contrast-to-noise ratio (m_CNR), CT number accuracy (m_CT#), and linearity (m_linearity).

All six metrics demonstrated a notable decline in IQ as the FOV_z increased from 2 to 27 cm. Specifically, the CNR decreased by half, while m_shading increased by 250 HU. The bowtie filter improved CT number accuracy at the periphery by approximately 100-140 HU, partially mitigating the impact of a larger FOV_z on IQ.

As the FOV_z increases, quantitative assessments reveal significant artifacts. Using a bowtie filter improves CNR and CT number accuracy while reducing shading and skin-line artifacts. For enhanced IQ in clinical therapy, minimizing the FOV_z is recommended. The evaluation framework established in this study provides a valuable tool for system comparison and assessing scatter correction techniques, aiding in accurate low-contrast detection and supporting advancements in online and adaptive radiotherapy.