Imaging of hepatic hemangiomas using multirotation dynamic SPECT

Hepatic hemangiomas are benign vascular lesions with 3 to 7% prevalence in the population, often appearing as suspicious masses during CT or ultrasound examination. Diagnosis through nuclear medicine relies heavily on the acquisition of a characteristic temporal uptake of Tc99m-labelled red blood cells in the lesion. Dynamic SPECT (dSPECT), a unique reconstruction method that generates a temporal series of 3D images, may potentially improve upon the 2D planar imaging currently used to capture dynamic behavior. A Monte-Carlo simulation of the digital NCAT phantom with a simulated liver hemangioma of increasing activity was used to generate 640 projections with and without scatter over 5 rotations of a dual-headed gamma camera. The dSPECT was used to reconstruct the 5-rotation projection data set, generating 80 frames in total. Also, the 5 rotations were divided into 5, 360° data sets and reconstructed with a static rescaled-block-iterative (RBI) method with and without attenuation correction for comparison. Clinical feasibility was also assessed. The clinical protocol consists of 1.5-min of planar dynamic imaging, four 5-min SPECTs, and a final 30-min SPECT with transmission imaging to provide a patient-specific attenuation map. The four 5-min SPECTs were reconstructed both as 4 separate static images using the RBI method with and without attenuation correction, and as a single dynamic series using the dSPECT algorithm in order to compare each technique´s ability to provide dynamic information. The dSPECT reproduces the simulated activity accurately. Regional TACs can be generated from multiple slow-rotation dynamic SPECT acquisitions and could be used in the diagnosis of hemangioma.