Improvement of image quality using interpolated projection data estimation method in SPECT

Objective General data acquisition for single photon emission computed tomography (SPECT) is performed in 90 or 60 directions, with a coarse pitch of approximately 4–6° for a rotation of 360° or 180°, using a gamma camera. No data between adjacent projections will be sampled under these circumstances. The aim of the study was to develop a method to improve SPECT image quality by generating lacking projection data through interpolation of data obtained with a coarse pitch such as 6°. Methods The projection data set at each individual degree in 360 directions was generated by a weighted average interpolation method from the projection data acquired with a coarse sampling angle (interpolated projection data estimation processing method, IPDE method). The IPDE method was applied to the numerical digital phantom data, actual phantom data and clinical brain data with Tc-99m ECD. All SPECT images were reconstructed by the filtered back-projection method and compared with the original SPECT images. Results The results confirmed that streak artifacts decreased by apparently increasing a sampling number in SPECT after interpolation and also improved S/N ratio of the root mean square uncertainty value. Furthermore, the normalized mean square error values, compared with standard images, had similar ones after interpolation. Moreover, the contrast and concentration ratios increased their effects after interpolation. Conclusion These results indicate that effective improvement of image quality can be expected with interpolation. Thus, image quality and the ability to depict images can be improved while maintaining the present acquisition time and image quality. In addition, this can be achieved more effectively than at present even if the acquisition time is reduced.