Performance of a SPECT collimator-detector response reconstruction algorithm: phantom studies and validation in inflammation clinical studies.

ABSTRACT

AIM: To evaluate evolution for bone (EfB), one of the recent corrective collimator detector response (CDR) image reconstruction packages, resolution recovery and iterative reconstruction protocols were compared to the standard Filtered Back Projection (FBP) protocol and the possibility of time reduced acquisition (15 instead of 30 seconds/projection) was evaluated. METHODS: Tomographic spatial resolution, contrast, noise and signal-to-noise ratio (SNR) were analyzed in phantom studies. Patient studies were conducted on sixteen [99mTc]HMPAO labeled leukocyte patients with suspected local inflammation, submitted to SPECT/CT and conventional leukocyte scintigraphy. Target background ratio (TBR) and boundary of each lesion were calculated. Images were also analyzed by visual inspection by two blinded expert physicians. RESULTS: a) Phantom resolution in phantom improves with the use of EfB mainly due to iterative reconstruction methods vs FBP. EfB shows a reduction in noise. No differences were found in contrast data, while SNR changes were mainly due to changes in noise. Time reduction, while increasing image noise, does not modify resolution; nevertheless, half-time EfB noise is the same as full-time FBP noise. b) PATIENTS lesion boundary is better defined in patient studies using iterative reconstruction vs FBP; no changes were observed for CDR vs iterative OSEM, or for full-time vs half-time acquisition. TBR is comparable in iterative and FBP protocols, while it is improved by EfB. At visual inspection, a higher score is always associated with EfB. CONCLUSION: Resolution, noise, SNR and TBR improve when applying the resolution recovery. EfB permits reduction of acquisition time without compromising image quality.