A comparison of different types of commercial FBP and OSEM SPECT reconstruction software

This study aimed at comparing the performance of filtered backprojection (FBP) and ordered subset expectation maximisation (OSEM) reconstruction algorithms available in several types of commercial SPECT software.
Methods: Numerical simulations of SPECT acquisitions of two phantoms were used: the NEMA line used for the assessment of SPECT resolution and a phantom with uniform, hot and cold rod compartments. The investigated types of software were: General Electric Xeleris and Vision, Philips Jetstream, Segami Mirage, Siemens eSoft and Icon, and Sopha Medical Vision XT. For FBP, no filtering and filtering of the projections with either Butterworth (order 3 or 6) or Hanning filters at various cut-off frequencies were considered. For OSEM, the number of subsets was 1, 4, 8, or 16 and the number of iterations was chosen to obtain a product number of iterations times the number of subsets equal to 16, 32, 48 or 64. The line phantom enabled us to obtain the reconstructed central, radial and tangential full-widths at half-maximum. The uniform compartment of the second phantom delivered the mean reconstructed pixel values and the standard deviations from which the coefficients of variation were calculated. Hot and cold contrasts were obtained from its rod compartments.
Results: For FBP, full-widths at half-maximum, mean pixel values, coefficients of variation and contrasts were almost software independent. The only exceptions were: smaller (0.5 mm) full-widths at half-maximum for Vision, larger mean pixel values for Vision and XT, and better contrasts for Vision and XT for some filtering conditions. For OSEM, full-widths at half-maximum differed between the different types of software from 0.1 to 2.5 mm but these were almost independent of the number of subsets or iterations. There was a high dependence of the mean pixel value on the type of software used and a moderate dependence of the coefficient of variation. The contrasts were almost software independent. Mean pixel value varied greatly with the number of iterations for Mirage and Vision, and the coefficient of variation increased with the number of iterations for all types of software. The mean pixel value, the coefficient of variation and the contrasts were almost constant for a fixed product number of iterations times the number of subsets, whatever the number of subsets or iterations.
Conclusion: Most of the types of software were equivalent for FBP or OSEM reconstruction. However, a few differences were observed for some types of software and should be considered when using them.