• 목록

1Jong Min Seung,  2Joong Hyun Kim, 1Hong Jae  Lee, 1Hyun Joo Kim, 1,2Jae Sung Lee, 1,2Dong Soo

Lee

1Dept. of Nuclear Medicine, Seoul National Univ. Hospital

2Dept. of Nuclear Medicine, Seoul National U

Objective

There has been an effort to develop collimator-detector response (CDR) compensation algorithms recently. Those algorithms  achieve higher contrast, better spatial resolution and shorter scan time than standard iterative reconstruction.  The aim of this study was to  evaluate one of the newest CDR compensation algorithms, Evolution (GE Healthcare, WI, USA).

Methods

GE Discovery NM/CT 670 (GE SPECT/CT,  USA) was used for phantom  studies.  NEMA SPECT Triple Line Source Phantom (Data Spectrum Corp.,  NC, USA) study was performed  with short and long radiuses for  resolution  comparison.  NEMA PET  IEC  Body Phantom (Data Spectrum Corp.) study was performed  for the comparison of contrast ratio and background  variability  between full time (30  sec/frame)   and  half  time   acquisition   (15  sec/frame).   All  quantitative   evaluations  were performed following NEMA NU 1-2007 standards. For qualitative evaluation, Hoffman 2-D Brain Phantom (Data Spectrum Corp.) and patients’ SPECT (thyroid, knee, and spine) images were compared. All images were reconstructed with and without  Evolution algorithm, respectively.

Results

The FWHM  was slightly  better  with Evolution  (14.7 mm) than without  (15.1 mm) for long  radius scan, however,  it  showed opposite  result (13.3 vs.  12.9 mm) for  short  radius scan. The contrast ratio was higher with Evolution than without  for both of full time (15.1%) and half time acquisition (23.0%). The background variability was slightly higher with Evolution than without, but it was not significant. In 2D Hoffman phantom  images, there was no significant difference between with and without   Evolution.  However,  patient  images  with  Evolution  showed  better  image  quality  than without.

Conclusion

The Evolution  presented  better  image  quality  than  standard  iterative  reconstruction  algorithm. Thus, we expect that  CDR compensation  algorithm  such as Evolution  in iterative  reconstruction will be useful for clinical applications.