[Normal database (NDB) of 123I-IMP brain perfusion SPECT examination is affected by statistical image analysis in the presence or absence of scatter correction and attenuation correction]

Abstract

Purpose: The purpose of our study is the establishment of normal database (NDB) from persons (aged 50-80 years) for 3D-SSP analysis of 123I-IMP brain perfusion SPECT image, and we analyzed whether the presence or absence of image correction methods, scatter correction (SC) and attenuation correction (AC), affects the created NDB and 3D-SSP analysis.

Materials and methods: We created NDB from 63 healthy volunteers (32 males and 31 females, aged 50-80 years), and calculated the coefficient of variation for each pixel from the mean value and standard deviation. Next, we compared the visual assessments of the standard deviation images by each image correction method, and the coefficient of variation of SEE analysis in each part of the brain. Furthermore, we examined frontotemporal dementia and healthy volunteers by 3D-SSP analysis, and evaluated the differences of Z-score in the presence or absence of image correction methods.

Results: In NDB, the coefficient of variation was the minimum when SC and AC were not applied and the maximum in periventricular and cerebellum when SC and AC were applied. In Z-score image of 3D-SSP analysis, Z-score of the low value was the maximum when SC and AC were not applied.

Conclusion: It was shown that the results from NDB coefficient of variation and 3D-SSP analysis were affected by the differences of image correction methods. It is important to understand the feature of imaging reconstruction conditions in your own facilities, and evaluate 3D-SSP analysis.