SPECT Filters Audit
The purpose of this audit was to identify differences in how manufacturers implement what are nominally described as the same filter during SPECT reconstruction using filtered back projection. For example, using a Butterworth filter, order 6, cut off 0.3 cycles per pixel does not necessarily produce the same results in all systems. In order to identify differences between systems, SPECT projection data comprising computer generated simulations of a simple test object were reconstructed using a range of specified filters on a variety of different computer systems. Results were analysed to identify the frequency response of the actual filter used and compared with the theoretical frequency response. To do this it was only necesssary to test two examples of each available nuclear medicine software and so, unlike other audits conducted by the group, it was not be necessary to invite participation from everywhere in the UK. Instead only a limited number of centres with access to a range of different computer systems were involved.
This audit, which was led by Richard Lawson, is now complete. The results were presented at the European Association of Nuclear Medicine Meeting in Birmingham in October 2011 and have been submitted for publication in a nuclear medicine journal.
As a result of this audit the NMSQG have produced some tools to aid users who wish to convert Butterworth filters from one nuclear medicine computer system to another. This are available on the “What’s your Filter” page of this website