Simultaneous fluoroscopy and nuclear imaging: Impact of collimator choice on nuclear image quality.

Objectives A real time, simultaneous fluoroscopic and nuclear imaging device, consisting of a c-arm combined with four gamma cameras, is currently being developed by our group. Image guided procedures could benefit from the availability of live hybrid images during interventions. The objective of this interventional device is to deliver quick insight into radionuclide distributions rather than providing images with the highest diagnostic image quality. To acquire gamma images overlapping with the x-ray images, gamma cameras are placed next to the x-ray tube. The gamma camera projections are used to reconstruct a virtual projection that is geometrically equal to the x-ray projection, resulting in an always overlapping hybrid image. This requires an intermediate three-dimensional reconstruction of the activity distribution from the four gamma camera projections. The gamma cameras can be equipped with parallel or pinhole collimators. The purpose of this simulation study was to evaluate the impact of collimator choice on nuclear image quality.

Methods Projections of two digital image quality phantoms were simulated including noise and resolution effects, with an acquisition time of 10 s and a total activity of 150 MBq. This could represent typical values for live guidance of a radio-embolization procedure. The first phantom consists of a homogeneous cylinder and is used to determine noise level. The second phantom consists of a cylinder with spheres of 2.0, 2.5, 3.0, 3.5 and 4 cm diameter, respectively, with foreground:background intensity 1:10 and is used to determine the contrast of the different sphere sizes. Simulations were performed for 3, 5 and 7 mm pinholes and three parallel hole collimators (low-energy-all-purpose (LEAP), low-energy-high-resolution (LEHR) and low-energy-ultra-high-resolution (LEUHR)). Intermediate reconstruction was performed with maximum likelihood expectation-maximisation (MLEM) with PSF modelling. For evaluation purposes, the ideal projection was also constructed. The ideal projection is defined as the projection obtained as if there was a gamma camera with a cone-beam collimator focussing on the x-ray focal spot at the same position as the x-ray detector, hence directly capable of acquiring gamma images using the x-ray geometry. For all cases, noise level and contrast of the virtual projection were determined and compared with the ideal projection. In addition, the artefact level was assessed.

Results The 3 mm pinhole collimator led to the highest noise level. The 7 mm pinhole had the lowest noise level and was even slightly better than the ideal projection. This can be attributed to the fact that in the hybrid device four gamma cameras are employed. For the smallest sphere, contrast was up to six times lower in the virtual projections than in the ideal projection. For the largest sphere, contrast was slightly (20%) lower. In parallel hole geometries, streak artefacts could be observed.

Conclusions As compared with the ideal projection geometry, some concessions were made with respect to contrast in the hybrid images. However, the hybrid device has the substantial advantage of enabling live, always overlapping hybrid images. Such images can be of great value in guiding oncological interventions. $$graphic_7B836144-42C6-40F9-A6B0-DE80F3D38791$$