Bone-window CT (BW-CT) enhances contrast by displaying fluids and soft tissues in a narrow range of mid-grey pixel intensities. This results in dense structures like cortical and trabecular bone appearing bright against a low-contrast soft tissue background. This contrast allows for a detailed assessment of the bone mineral matrix, minimising interference from surrounding soft tissues.

We employed low flip angle (3°) PD-weighted radial StarVIBE and Cartesian VIBE sequences to replicate BW-CT contrast. These low flip angles produce images where fluids and soft tissues are mid-grey, and bone appears black (often termed "black bone" imaging) [1]. By inverting the grayscale of these images, we can then display bone as white, effectively mimicking the appearance of BW-CT.

The primary challenge with this approach is the low signal intensity inherent to the 3° flip angle, leading to a poor signal-to-noise ratio (SNR) and noisy images, especially centrally. Increasing the flip angle would improve SNR but undesirably enhance soft tissue contrast, compromising the desired bone-specific visualisation.

We iteratively developed the sequence through a service development process to optimise the balance between acquisition time (TA), SNR, slice resolution, and in-plane resolution. We explored Cartesian and radial k-space filling schemes, evaluating spatial resolution, artefacts, signal and contrast-to-noise ratios, and diagnostic utility for each. We implemented a deep learning (DL) VIBE sequence to address the low SNR issue. This technique uses trained neural networks to reconstruct images from undersampled k-space data and subsequently applies neural network-based image interpolation. For this purpose, Siemens Healthineers provided a work-in-progress Deep Resolve Boost and Deep Resolve Sharp VIBE research sequence.

Our initial evaluation focused on a single-station 3D volume (35 cm z-axis) covering the pelvis and lower lumbar spine. This region was selected due to its high bone and marrow content and its frequent involvement in metastatic cancer. The optimised techniques were then adapted for whole-body applications.