Clinical impact of nonuniform CT-based attenuation correction in brain perfusion SPECT/CT using (99m)Tc-ECD.

INTRODUCTION

Brain perfusion SPECT is commonly used to evaluate patients with cognitive impairments. Physical limits such as attenuation compromise image quality do not allow the most accurate depiction of radionuclide distribution, and thus, application of attenuation correction (AC) has been recommended. Some reports have demonstrated discordances between the uniform (UAC) and nonuniform CT-based correction (NUAC) procedures. The aim was to study the impact of these discordances on visual interpretation and their concordance with clinical symptoms.

METHODS

Twelve patients presenting cognitive impairments were included. Brain perfusion SPECT images were reconstructed using 2 AC methods. Qualitative image assessment was performed as uptake analysis in 21 predefined cortical ROIs for each patient. Interpretation of perfusion patterns was based on a 2-score uptake scale (normal and reduced/pathologic). Variation of uptake scores in UAC- versus NUAC-processed images and their concordance with clinical symptoms were studied.

RESULTS

Normal image patterns generated by UAC and NUAC methods were found in 226 (90%) of 252 and in 201 (80%) of 252 ROIs, respectively. No difference between UAC and NUAC methods was found in posterior brain areas. However, differences were recorded in 51 (20%) of 252 ROIs, and this discordance was located in the anterior areas (frontal and temporal lobes), and evaluation changed from normal to pathological patterns using NUAC method. Two years later, patients showing frontal hypoperfusion on NUAC brain SPECT images expressed clinical frontal lobe dysfunctions.

CONCLUSIONS

Discordances between UAC- and NUAC-processed images impact visual analysis of brain perfusion SPECT images. The NUAC-processed images show a good concordance with clinical symptoms, suggesting that it is an accurate method to correct attenuation.